Title: Improving personalized causal inference with information borrowed from heterogeneous data sources
Authors: Xiaoqing Tan - University of Pittsburgh (United States) [presenting]
Lu Tang - University of Pittsburgh (United States)
Abstract: Individualized causal inference, ranging from personalized medicine to customized marketing advertisement, has remained a hot topic. However, due to the limited sample size in a single study, estimating treatment effects or optimal treatment rules is often challenging. We propose a tree-based model averaging framework to improve the estimation efficiency of conditional average treatment effects (CATE) and optimal decision rules concerning the population of a targeted research site by leveraging models derived from potentially heterogeneous populations of other sites, but without them sharing individual-level data. To our best knowledge, there is no established model averaging approach for distributed data with a focus on improving the estimation of treatment effects. Drawing on a multi-hospital electronic health records network, we develop an efficient and interpretable tree-based ensemble of personalized treatment effect estimators to join results across hospital sites, while actively modeling for the heterogeneity in data sources through site partitioning. The efficiency of this approach is demonstrated by a study of causal effects of oxygen saturation on hospital mortality and backed up by comprehensive numerical results.